Search

Background: People with Down syndrome (DS) are at ultra-high risk of developing Alzheimer's dementia (AD) as they get older. This is primarily due to the triplication of the amyloid precursor protein gene (APP) on chromosome 21. Triplication of APP induces dysregulation of amyloid processing pathways, leading to increased lifelong over-production of amyloid-beta (Aβ) protein which aggregates to form plaques in cerebral tissues. Aβ can also be deposited in blood vessel walls leading to cerebral amyloid angiopathy (CAA). Plaque formation is accompanied by the development of neurofibrillary tangles consisting of tau proteins, neuroinflammation, synaptic injury and tissue atrophy. Over-time this contributes to progressive development of the clinical symptoms of dementia, particularly short-term memory difficulties, decline in functional skills, loss of independence, and eventually death. Despite people with DS sharing similar genetic risks for the development of AD, there are large variations in the age of onset. Whilst the current average age at diagnosis is 55 years, diagnosis before age 40 and over 65 years have been reported. In sporadic AD, poor cardiovascular health has been associated with increased risk of developing vascular dementia and AD, particularly when the risks are present in mid-life. It has not been established whether the same risks apply to people with DS. It was hypothesised that cardiovascular health and lifestyle factors may affect dementia onset and progression in people with DS, modulating the genetic risks and pattern of neuropathology. The aims of the study were: 1) To establish the cardiovascular profile in people with DS - including blood pressure, body mass index, lipid and glucose measurements - throughout adulthood and understand whether this differs from the general population; 2) To determine whether cardiovascular health in mid-life affects incidence of dementia in people with DS; 3) To explore the relationship between cardiovascular health parameters with cognitive outcomes and plasma biomarkers as a way to explore potential mechanistic pathways which link cardiovascular health and dementia pathology; 4) To consider targets for monitoring and therapeutic interventions which may help to modify the risk of dementia incidence in people with DS. Methods: Two complementary approaches were used in a synergistic manner. These were a population based cohort study utilising data from UK-based primary care electronic health records, and a deeply phenotyped cohort study which collected health, cognitive and biomarker data from adults with Down syndrome approximately every 2 years (the LonDowns cohort). Data were extracted from the Clinical Practice Research Datalink (CPRD) for all people with DS identified in the dataset, alongside controls matched on age, sex and GP practice on a ratio up to 4:1. The incidence rate ratios of common cardiovascular co-morbidities were calculated between people with DS and controls. Direct measurements of nine cardiovascular parameters (CVPs) were used to analyse patterns across adulthood. These were systolic blood pressure (SBP), diastolic blood pressure (DBP), pulse pressure (PP), body mass index (BMI), total cholesterol (TC), high and low density lipoproteins (HDL and LDL), total triglycerides (Tg), and glycated haemoglobin (HbA1c). Results were presented graphically using scatterplots. Regression models were constructed to examine the effects of covariates (age, sex, level of intellectual disability). Multivariate analyses were used to compare differences between five-year age bands. Following this, survival curves were fitted using mid-life cardiovascular parameters as the predictor variables and incidence of dementia as the outcome. Age, sex, level of intellectual disability, history of congenital heart disease (CHD), cardiovascular disease diagnoses, hypothyroidism and diabetes (both type 1 and 2) were included as covariates in survival models. Health, cognitive and plasma biomarker data from the LonDowns cohort were used to explore interactions between CVPs, cognitive outcomes, and markers of amyloid-beta and neuropathology. The focus was on blood pressure, BMI and exercise levels given their relationship with dementia incidence or symptomatology in earlier analyses. The cognitive outcomes used were paired associate learning (PAL) and semantic verbal fluency (VF) tasks alongside the CAMDEX-DS and short adaptative behaviour scale (SABS) informant questionnaires. Established plasma biomarkers relating to AD-neuropathology were analysed: Aβ40, Aβ42, Aβ42/40 ratio, neurofilament light (NfL), glial fibrillary acidic protein (GFAP), and phosphorylated-tau181 (p-tau181). Eight complement components were additionally examined (C3, C4, C1inh, C1q, CR1, clusterin, FH and FI). Two-sided Pearson correlations were calculated between the measures before the sample was split into an older and younger (below age 35 years) cohort. Regression models were constructed to evaluate the relationship between CVPs (dependent factors) and cognitive and biomarker values (independent factors), with adjustments for age, sex, and level of intellectual disability. These were performed cross-sectionally, and longitudinally in a sub-group of participants who had additional data available from approximately five years earlier. Comparisons were made between people with and without AD using two-way independent t-tests and regression models. Finally, principal component analysis was performed to assess the relationships between factors. Significance was set at p < 0.05 throughout. Results: 6,424 adults with DS were identified in the CPRD, matched to 23,143 controls. People with DS had a distinct cardiovascular health profile throughout adulthood compared to controls. They had significantly lower SBP, DBP and PP values. Minimal increases with age were observed. People with DS had higher BMI, and this was particularly in younger adulthood. A decline in BMI was observed after age 40 years in people with DS, which was steeper than that observed in controls. Women with DS had higher mean BP and BMI than men, whereas in controls men had higher average BP and BMI values. TC, LDL and HDL patterns were similar in both groups, but the DS cohort had lower mean TC and HDL. Tg and HbA1c patterns were different. Tg was more homogenous across ages compared to controls, with lower values observed in older people with DS. Trends of HbA1c with age were similar in people with diabetes in both DS and controls, although the DS cohort had slightly higher mean HbA1c values suggestive of poorer glycaemic control. In people without diabetes there was a plateauing of HbA1c values with age in DS but not in controls, and mean HbA1c values were lower in DS than controls. Survival analyses showed increased incidence of AD in people with low PP (< 40mmHg), Hazard Ratio (HR) 1.45; low DBP (< 62mmHg), HR 1.77, and low SBP (< 100mmHg), HR 1.83. High SBP (>140mmHg), HR 2.04 also increased risk but there were few people who experienced these higher values. Raised BMI >25kg/m2 increased incidence dementia risk (HR 1.63) and the risk was the same whether someone was overweight (25-30kg/m2) or obese (BMI >30kg/m2). Hypothyroidism, history of CHD, and male sex were significant contributors in the survival models. There was a cumulative effect of risk. People with 1 or 2 risk factors had HRs of 3.15 compared to those with zero risks, whilst people with 3 or more cardiovascular risk factors had HRs of 5.69. In the LonDowns cohort (n=92), NfL, p-tau181 and GFAP were negatively correlated with cognitive outcomes and positively correlated with dementia symptom burden. In the older cohort BMI was negatively related to GFAP and NfL, and positively with PAL scores. People engaged in higher levels of exercise had lower CAMDEX-DS scores indicating less dementia symptomatology. Low SBP was associated with higher levels of NfL. People with greater numbers of comorbidities had higher NfL and GFAP levels, with lower Aβ42/Aβ40 ratios. This was alongside higher CAMDEX-DS scores and poorer PAL performance. In the younger cohort SBP was negatively associated with NfL and GFAP levels. PP was positively associated with p-tau181. People with increased numbers of cardiovascular co-morbidities had higher NfL and lower Aβ40 levels. At a whole group level, BMI was positively associated with C3, FH and FI. People with dementia had higher GFAP, NfL, p-tau181, and C1q levels, with lower FI, clusterin and C4. In principal component analyses, NfL, GFAP and p-tau181 corresponded with CAMDEX-DS scores, and were good predictors of dementia status. Blood pressure measures grouped with Aβ measures but not NfL, GFAP or p-tau181. BMI did not correspond with the other CVPs, but was related to complement components and Aβ42/40 ratio. Discussion: People with DS have a distinct cardiovascular profile and this is present throughout adulthood. Blood pressure, BMI, and glucose homeostasis appear to be particularly affected. Additional interactions were observed on lipid levels, with lower HDL and TC levels in people with DS. These differences may be attributable to intrinsic autonomic dysfunction, direct and downstream genetic effects, atherosclerotic mechanisms, alterations in metabolic pathways, pro-inflammatory states, or social and environmental factors.